New research shows how small video-language models can gain advanced reasoning through synthetic chain-of-thought distillation and difficulty-aware fine-tuning, closing the gap with large models at a fraction of the compute cost.
Open-weight Chinese models from labs like DeepSeek and Alibaba are rapidly narrowing the performance gap with Silicon Valley's closed frontier systems—reshaping the economics of AI and the tools that power synthetic media generation.
Sakana AI researchers unveil PC-ALM, a layer-local training method that replaces global backpropagation and successfully trains 1000-layer networks, hinting at more scalable and parallelizable deep learning.
Deepfake detection systems have seen accuracy fall by 49% over the past year as AI video generation tools grow more sophisticated, widening the gap between synthetic media creation and the tools built to catch it.
Speculative decoding lets AI models predict multiple tokens ahead using a small draft model verified by a larger one, dramatically cutting inference latency without sacrificing output quality — a technique reshaping how fast generative models run.
Nvidia is reportedly weighing an investment in Anthropic's upcoming IPO, a move that would further intertwine the world's dominant AI chipmaker with one of the leading frontier model labs behind Claude.
Hancom With's SPEEKEY voice technology placed second in a global challenge for voice deepfake creation and detection, underscoring rising momentum in synthetic audio authentication research.
A new arXiv paper dissects how LoRA rank selection impacts quality, memory, and generalization when fine-tuning diffusion models — a key technique behind custom image and video generation for synthetic media.
DeepSeek's new V4.1-Flash pushes a 1M-token context window with FP4 KV cache compression and cross-layer attention reuse, slashing memory costs for long-context inference and reshaping the efficiency frontier for large models.
Autoregressive language models generate text token by token. A new wave of diffusion-based LLMs aims to produce entire passages in parallel, promising dramatic speedups and rethinking how synthetic text is created.
Suno is replacing its AI music generation models with a new system trained on licensed catalogs as copyright lawsuits from major labels pile up—a pivotal moment for synthetic audio and the data provenance debate.
A technical deep dive into attention mechanisms, the mathematical core of transformer architectures that power today's generative AI systems from language models to video and image synthesis tools.